werkr Documentation

werkr is an AI-native developer environment manager that allows you to configure and create environments, sandboxes, and AI workspaces in a fast and reproducible way.

What is werkr?

werkr provides deterministic execution environments. It gives you a single, declarative way to define where and how code runs - whether it's a developer on a laptop, a CI pipeline, or an AI agent in a sandbox.

Key Features

  • Reproducible - Define environments once, spin them up identically everywhere.
  • Local or cloud - Run on a developer laptop, in a container, or in the cloud.
  • Deployable in your tenant - Your data, your rules, your infrastructure.
  • Configurable - Define defaults while allowing local customizations.
  • AI-native - Purpose-built for AI tooling and agents.
  • Built for human and machine code - Same guardrails for both.

How it works

werkr uses a Kubernetes-inspired declarative model. You define Workstations and Workstation Configurations as YAML manifests. werkr reconciles these definitions against the actual state of your infrastructure, creating and configuring environments automatically. 

# my-workstation.yaml
# Supported OS: ubuntu-24.04, debian-13
apiVersion: werkr.dev/v1alpha1
kind: Workstation
metadata:
  name: my-dev-env
  labels:
    team: platform
spec:
  os:
    id: ubuntu-24.04
  provider: aws
  providerConfig:
    region: us-east-1
    machineType: t2.micro
  resources:
    diskSize: 20
---
# The WorkstationConfig matches workstations by label.
# Here, any workstation with "team: platform" gets this config applied.
apiVersion: werkr.dev/v1alpha1
kind: WorkstationConfig
metadata:
  name: platform-tools
spec:
  selector:
    matchLabels:
      team: platform
  portForward:
    autoForward: true
  contents:
    devtools:
      - docker
      - kubectl
      - kind
    packages:
      - tmux

Apply the manifest to create and start the workstation as a remote EC2 instance in AWS: 

wer apply -f my-workstation.yaml

werkr automates SSH access to the machine and sets up port-forwarding, so that things like OAuth authentication flows work flawlessly between your local machine and the remote environment.

Next Step

Ready to try it? Start with the tutorial to go from zero to a running workstation in a few minutes. If you've already installed werkr, skip ahead to the installation guide for reference.